Human-Object Interaction (HOI) detection remains challenging due to the semantic ambiguity of interaction categories and the limited discriminability of their feature representations. Existing approaches often improve recognition by employing sophisticated models or auxiliary textual annotations. While effective in certain gains, these solutions incur additional computational or annotation costs and struggle to capture intrinsic interaction regularities. To address these issues, we propose Pose-Guided Multi-Cue Explicit Query Construction (PM-EQC), a unified Transformer-based framework that builds upon collaborative modeling of appearance, spatial, and pose cues for discriminative interaction reasoning. At its core, the Collaborative Multi-Cue Query Constructor (CM-CQC) jointly models dependencies among visual cues to generate explicit query embeddings. CM-CQC further incorporates a hierarchical pose contextualization mechanism: global body configurations adaptively guide attention to local critical joints, yielding fine-grained pose embeddings and more precise interaction disambiguation. Owing to its modular design, PM-EQC integrates seamlessly with diverse backbones and benefits from their advances. Extensive experiments on PhysLab, HICO-DET, and V-COCO datasets demonstrate that PM-EQC achieves state-of-the-art performance, and the code is publicly available at https://github.com/ZMHSDUST/ PM-EQC.

The recommendation based on dynamic graph sequences aims to reveal complex evolutionary patterns in user-item interactions. Existing methods make predictions by encoding attribute contents through similarity but lack dynamic modeling of fine-grained attribute scenarios, resulting in a deviation in user interest representation. To address above issues, we propose a novel Dynamic Graph multi-granularity Attribute Scene evolution sequence Recommendation (DGASR) to enhance content features and reduce user interest bias by finer granularity modeling dynamic attributes. Firstly, we design an attribute-aware reconstruction module to model attribute interest distribution to reconstruct attributes and graphs. Subsequently, we design an attribute-aware long-short term module. It enhances long-term evolution characteristics of user behavior under attribute scene changes and constrains the consistency of users’ short-term interest distribution, achieving dynamic modeling of behavioral preferences under attribute scene distribution. Finally, DGASR achieves state-of-the-art results on three benchmark datasets, significantly outperforming several typical cold-start methods.

Heterogeneous Graph Neural Networks(HGNNs) are widely regarded as an effective tool for modeling data with graph structures in recommendation. Current research lacks modeling of user attribute and project attribute distribution preferences, limiting graph structure optimization potential. In response to these challenges, we propose a novel Heterogeneous Graph Dual-structure Optimization based Attribute-aware for Recommendation systems (HDSAR). It captures users’ personalized preferences through attribute-aware enhancement and uses dual-structure optimization to improve recommendation performance. First, we design an attribute-aware enhancement module to significantly enhance the relevance of attributes between users and items. Second, we use attribute-aware signals to explicitly filter heterogeneous neighbor ranges to preserve high- quality structural neighborhoods. Then, we employ contrastive learning to enhance the consistency of attribute-aware signals and heterogeneous structures to implicitly optimize the structural learning. Experiments on two real-world datasets demonstrate that HDSAR’s recommendation performance surpasses that of state-of-the-art methods.

ABSTRACT Pedestrian detection is crucial in agricultural environments to ensure the safe operation of intelligent machinery. In orchards, pedestrians exhibit unpredictable behavior and can pose significant challenges to navigation and operation. This demands reliable detection technologies that ensures safety while addressing the unique challenges of orchard environments, such as dense foliage, uneven terrain, and varying lighting conditions. To address this, we propose ReB‐DINO, a robust and accurate orchard pedestrian detection model based on an improved DINO. Initially, we improve the feature extraction module of DINO using structural re‐parameterization, enhancing accuracy and speed of the model during training and inference decoupling. In addition, a progressive feature fusion module is employed to fuse the extracted features and improve model accuracy. Finally, the network incorporates a convolutional block attention mechanism and an improved loss function to improve pedestrian detection rates. The experimental results demonstrate a 1.6% improvement in Recall on the NREC dataset compared to the baseline. Moreover, the results show a 4.2% improvement in and the number of parameters decreases by 40.2% compared to the original DINO. In the PiFO dataset, the with a threshold of 0.5 reaches 99.4%, demonstrating high detection accuracy in realistic scenarios. Therefore, our model enhances both detection accuracy and real‐time object detection capabilities in apple orchards, maintaining a lightweight attributes, surpassing mainstream object detection models.

ABSTRACT In recent years, self‐attention mechanisms have demonstrated remarkable performance across various computer vision tasks, gradually emerging as a mainstream approach. However, compared to traditional Convolutional Neural Networks (CNNs), its high quadratic complexity and limited adaptability to 2D structures have constrained its broader application and adoption. To enhance the feature extraction capability of CNNs, this work focuses on augmenting input information and introduces a novel architectural unit called the Global Query Vector (GQ Vector). The proposed unit adopts a co‐evolutionary architecture consisting of a parallel branch and the main backbone network, which continuously integrates and refines global semantic information during forward propagation, establishing a cross‐layer, persistent context memory mechanism. This design enables progressive accumulation and refinement of contextual information, thereby enhancing the CNN's capacity to model long‐range dependencies. Building on this, we propose a novel CNN architecture named a Global Query Convolutional Network (GQConvNet). It can be seamlessly integrated into existing CNN frameworks, further enhancing their performance. For example, on the ImageNet‐1K dataset, a ResNet‐50 model augmented with GQ Vector achieves a 1.7% improvement in Top‐1 accuracy over baseline models. This work offers a fresh perspective on optimizing CNNs, with substantial academic value and practical implications.

ABSTRACT May‐Happen‐in‐Parallel (MHP) analysis serves as the basis for many concurrency bugs analyses. Inadequate handling of the coupling between locks and thread creation statements, as well as inter‐procedural locks, can lead to a loss of precision in MHP analysis. To address these issues, this paper proposes a new MHP analysis for C that operates at the segment granularity rather than individual statements. By constructing a Segmented Thread‐sensitive Control Flow Graph (STCFG) for a program, statements are grouped into different segments. Context information is added to these segments to capture the semantics of Pthreads operations, thereby identifying Happens‐Before (HB) and conflict relationships between segments. To compute MHP information for statement pairs, it is sufficient to examine the relationship between segments to infer the relationship between statements. We implement our algorithm in LLVM and evaluate it using eight test cases as well as four programs from the SPLASH2 benchmark suite. Preliminary results show that our method provides higher precision and achieves higher efficiency.

The dynamic topologies and sensitive information exchanged among autonomous vehicle groups make them prime targets for attackers. In particular, in a collusive attack scenario, malicious nodes can collaborate to manipulate the trust evaluation system, thereby compromising the security of the entire vehicle group. To handle this limitation, this work proposes a privacy-preserving method for forming autonomous vehicle groups in a collusive attack scenario. First, we introduce a distributed trust evaluation algorithm based on a federated learning topology, which preserves local data privacy while facilitating reliable inter-vehicle trust computation. Then, we propose a PageRank-based detection mechanism that analyzes the trust propagation network to identify potential collusive attackers. Finally, we present a privacy-preserving method for autonomous vehicle group formation. Experimental results show that our proposed approach significantly improves the security and stability of autonomous vehicle groups compared to existing methods.

In the upcoming 6G era, the demand for network communication and computing is expected to surge and diversify. A Space-Air-Ground Integrated Network (SAGIN) is introduced as a solution to provide seamless global connectivity. Meanwhile, network slicing technology can further enhance network capabilities by supporting customized services. However, the joint optimization of multi-dimensional resource allocation and task offloading decisions presents a significant challenge in dynamic and complex environments with diverse task types. In this work, we establish an SAG IN paradigm that integrates network slicing. A multi-level optimization scheme for resource allocation and task offloading is proposed to improve Key Performance Indicators (KPIs) and Quality of Service (QoS). Specifically, we propose a State Encoding-based Multi-Agent Soft Actor-Critic algorithm (SE- MASAC) within a Centralized Training with Decentralized Execution (CTDE) architecture. This algorithm processes current sensing states and historical information, leveraging reinforcement learning to make inter-slice resource and offloading scheduling decisions. Intra-slice decisions for intelligent User Equipments (iUEs) are determined based on their sensing data-driven priority. Simulation results demonstrate that our approach outperforms baselines in terms of system utility, QoS satisfaction, and various KPIs.

Forming a stable autonomous vehicle group is extremely challenging in an urban scene, which is disturbed by many environmental factors, e.g., manned vehicles, roadside obstacles, traffic lights, and pedestrians. Existing work focuses on autonomous vehicle group formation (AVGF) in a highway scene only. Its outcomes cannot be directly applied to an urban scene because of different environmental factors and poor communication quality. This work presents an autonomous vehicle group model in an urban scene. First, it proposes a prediction method to analyze the impact of environmental factors on communications among autonomous vehicles. Then, it defines preperception degree, vehicle activity, and mobility similarity of autonomous vehicles and selects leader vehicles based on them. Next, it measures connectivity, coupling, and timeliness increments of a vehicle group, based on which a vehicle group model is formulated. Finally, it solves the proposed vehicle group model by using a modified distributed multiobjective optimization method, proves its convergence, and analyzes its time complexity. The simulation results on synthetic and real roads show that the proposed prediction method achieves lower errors than XGBoost and a multilayer perceptron, and the proposed vehicle group model outperforms two AVGF methods and a dynamic clustering method for vehicular ad-hoc network.

Formulating a cooperative autonomous vehicle group is challenging in an urban scene that has complex road networks and diverse disturbance. Existing methods of vehicle cluster cooperation in a vehicular ad-hoc network cannot be applied to autonomous vehicles because the latter have different requirements for a vehicle group structure and communication quality. Existing studies focus on autonomous vehicle group cooperation in closed and highway scenes only. Their outcomes cannot be directly applied to an urban scene because of its complex road conditions, incomplete cooperation properties, and lack of a vehicle group size control strategy. In this work, we formulate a cooperation model for autonomous vehicle groups in such scene. First, we analyze cooperation criteria based on the non-colliding aggregate motion of flocks and deduce the connectivity, coupling, timeliness, evolvability, and adaptivity of a vehicle group, based on which we propose a cooperation model. Next, we solve our model by using a modified distributed evolutionary multi-objective optimization method, prove its convergence, and analyze its computational complexity. Finally, we conduct simulations on synthetic and real roads to show its performance in terms of average connectivity, coupling, timeliness, evolvability, and adaptivity of vehicle groups.

Accurately processing dynamic evolution events is extremely challenging for autonomous vehicle groups in an urban scene, which can be disturbed by manned vehicles, roadside obstacles, traffic lights, and pedestrians. Existing work focuses on a dynamic evolution method for such groups in a highway scene only. Its outcomes cannot be directly used to an urban scene due to different environmental factors, incomplete dynamic evolution events, and lack of simulation evaluation with real road networks. In this work, we present a dynamic evolution method for such groups in an urban scene. First, we analyze their dynamic evolution reasons. Then, we abstract five dynamic evolution events, i.e., joining, leaving, merging, splitting, and disappearing, and introduce a dynamic evolution method to process them. Finally, we deduce the evolvability that can reflect dynamic evolution states of a vehicle group. The simulation results in synthetic and real urban scenes show that the connectivity, coupling, timeliness, and evolvability of vehicle groups using the proposed dynamic evolution method are higher than those of using a dynamic evolution method for a highway scene.

Forming a stable autonomous vehicle group is extremely challenging in a highway scene that has several entrances and exits. Existing studies focus on centralized autonomous vehicle groups with leading nodes. Such groups suffer from unbalanced computing tasks, asymmetric information, and weak stability. This article introduces a side chain consensus-based decentralized autonomous vehicle group formation method in a highway scene. First, we side chain consensus to describe states of autonomous vehicles. Then, we give decentralized autonomous vehicle group formation and maintenance methods based on side chain consensus. Finally, we conduct simulations to evaluate the quality of side chain consensus and stability of vehicle groups, which shows that our method has better properties in the balance of computing tasks, information symmetry, and stability than existing methods.

Accurate estimation of network capacity is very important for Vehicular Infrastructure-based NETwork (VINET) in an urban scene that may involve greatly dynamic typology and complex driving conditions. The node mobility, network behavior, and network scale of a VINET are different from those of a wireless network, and, therefore, the existing capacity estimation methods of wireless networks cannot be used to estimate VINET capacity. In addition, most existing studies on VINET capacity only derive asymptotic descriptions when the number of nodes is large enough. In this work, a novel approach is proposed for the modeling and calculating VINET capacity. More specifically, we first analyze communication characteristics in a VINET, and introduce two transmission modes, i.e., a vehicle-based mode and a Road Side Unit (RSU)-based one. Then, we propose a probability-based transmission mode selecting strategy with which vehicle nodes can choose either transmission mode independently and such choice is probabilistic. Next, we analyze the characteristics of an RSU-based mode, divide a VINET into a number of communities according to the position and communication range of RSUs, and derive the capacity contributed by an RSU-based mode. Then, we calculate the capacity contributed by a vehicle-based mode based on fluid mechanics. Finally, the VINET capacity can be calculated. The proposed VINET capacity estimation approach is validated to be consistent with simulation results.

Maintaining network connectivity is an important challenge for vehicular ad hoc network (VANET) in an urban scene, which has more complex road conditions than highways and suburban areas. Most existing studies analyze end-to-end connectivity probability under a certain node distribution model, and reveal the relationship among network connectivity, node density, and a communication range. Because of various influencing factors and changing communication states, most of their results are not applicable to VANET in an urban scene. In this article, we propose a connectivity prediction-based dynamic clustering (DC) model for VANET in an urban scene. First, we introduce a connectivity prediction method (CP) according to the features of a vehicle node and relative features among vehicle nodes. Then, we formulate a DC model based on connectivity among vehicle nodes and vehicle node density. Finally, we present a DC model-based routing method to realize stable communications among vehicle nodes. The experimental results show that the proposed CP can achieve a lower error rate than the geographic routing based on predictive locations and multilayer perceptron. The proposed routing method can achieve lower end-to-end latency and higher delivery rate than the greedy perimeter stateless routing and modified distributed and mobility-adaptive clustering-based methods.

Accurately estimated data transmission ability is important in operating a vehicular ad-hoc network (VANET), which has limited bandwidth and highly dynamic typology. The mobility behavior of traditional wireless networks is different from VANET's, and existing results on the former are not applicable to VANET directly. Most existing studies on VANET capacity estimation focus on asymptotic descriptions. In them, messages sent and received by vehicle nodes are composed of data packets, and vehicle nodes can move along roads only. In this paper, a modeling and calculation approach for accurate VANET capacity is proposed. We transfer vehicle nodes to data packets and then abstract data packets that can move along roads into data flow in virtual pipelines. Then, we derive a fluid mechanics-based data flow model and propose capacity calculation equations. According to network scale, network capacity is divided into following three stages: linear growth, maintenance, and decline. This paper demonstrates that the data flow model-based capacity is consistent with that of simulation results.